Update app.py

app.py

@@ -56,8 +56,8 @@ EXAMPLE_SEQUENCES = [
     ],
 ]
 
-MAX_MAX_NEW_TOKENS = 256
-DEFAULT_MAX_NEW_TOKENS = 100
+MAX_MAX_NEW_TOKENS = 1024
+DEFAULT_MAX_NEW_TOKENS = 512
 
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
@@ -90,7 +90,7 @@ def stream_response(
     temperature: float = 0.6,
     top_p: float = 0.9,
     top_k: int = 50,
-    repetition_penalty: float = 1.2,
+    repetition_penalty: float = 1.0,
 ) -> Iterator[str]:
 
 
@@ -240,24 +240,24 @@ with gr.Blocks(theme=theme, css_paths="style.css", fill_height=True) as demo:
                 do_sample_checkbox = gr.Checkbox(label="Enable sampling", value=False)
                 temperature_slider = gr.Slider(
                     label="Temperature",
-                    minimum=0.1,
+                    minimum=0.0,
                     maximum=4.0,
                     step=0.1,
-                    value=0.6,
+                    value=0.0,
                 )
                 top_p_slider = gr.Slider(
                     label="Top-p (nucleus sampling)",
                     minimum=0.05,
                     maximum=1.0,
                     step=0.05,
-                    value=0.9,
+                    value=1.0,
                 )
                 top_k_slider = gr.Slider(
                     label="Top-k",
                     minimum=1,
                     maximum=1000,
                     step=1,
-                    value=50,
+                    value=1,
                 )
                 repetition_penalty_slider = gr.Slider(
                     label="Repetition penalty",